An m-Health intervention to improve education, self-management, and outcomes in patients admitted for acute decompensated heart failure: barriers to effective implementation
Georgios Zisis, M. Carrington, Brian Oldenburg, Kristyn Whitmore, Maria Lay, Quan Huynh, Christopher Neil, Jocasta Ball, Thomas H. Marwick
Abstract
Abstract Aims Effective and efficient education and patient engagement are fundamental to improve health outcomes in heart failure (HF). The use of artificial intelligence (AI) to enable more effective delivery of education is becoming more widespread for a range of chronic conditions. We sought to determine whether an avatar-based HF-app could improve outcomes by enhancing HF knowledge and improving patient quality of life and self-care behaviour. Methods and results In a randomized controlled trial of patients admitted for acute decompensated HF (ADHF), patients at high risk (≥33%) for 30-day hospital readmission and/or death were randomized to usual care or training with the HF-app. From August 2019 up until December 2020, 200 patients admitted to the hospital for ADHF were enrolled in the Risk-HF study. Of the 72 at high-risk, 36 (25 men; median age 81.5 years; 9.5 years of education; 15 in NYHA Class III at discharge) were randomized into the intervention arm and were offered education involving an HF-app. Whilst 26 (72%) could not use the HF-app, younger patients [odds ratio (OR) 0.89, 95% confidence interval (CI) 0.82–0.97; P < 0.01] and those with a higher education level (OR 1.58, 95% CI 1.09–2.28; P = 0.03) were more likely to enrol. Of those enrolled, only 2 of 10 patients engaged and completed ≥70% of the program, and 6 of the remaining 8 who did not engage were readmitted. Conclusions Although AI-based education is promising in chronic conditions, our study provides a note of caution about the barriers to enrolment in critically ill, post-acute, and elderly patients.